What does Signal actually check?

Browsers flag non-secure sites and Google ranks them lower. Visitors who see a security warning leave before they ever read your offer.

If your homepage returns an error code, search engines and AI crawlers drop it from their results entirely.

Without a canonical tag, the same page can be indexed under several addresses, splitting your ranking strength between them.

A noindex tag tells Google to remove the page from search. On a homepage it makes the business effectively invisible.

robots.txt is the first file every crawler reads. A missing or broken one leaves crawler behaviour to chance.

A sitemap is the menu you hand to Google. Without it, new and deep pages can take weeks longer to be discovered.

Every redirect hop adds delay and leaks ranking signal. Chains of 2+ hops slow both people and crawlers down.

Without a modern doctype browsers fall back to quirks mode, which breaks layout and parsing in subtle ways.

A missing charset can garble special characters and forces the browser to guess the encoding.

GPTBot feeds ChatGPT training and browsing. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

OAI-SearchBot feeds ChatGPT search results. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

PerplexityBot feeds Perplexity answers. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

ClaudeBot feeds Claude by Anthropic. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

Google-Extended feeds Gemini and Google AI features. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

Applebot-Extended feeds Apple Intelligence and Siri. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

Bingbot feeds Bing search and Microsoft Copilot. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

CCBot feeds Common Crawl, used by many AI models. If it is blocked, your business cannot appear when customers ask that assistant for recommendations.

llms.txt is an emerging standard that hands AI assistants a curated summary of your site — an easy head start over competitors who lack one.

Structured data is how you state facts about your business in a format machines trust. Google rich results and AI answers both draw on it.

Broken JSON-LD is worse than none: crawlers silently discard it, and you lose the rich results you thought you had.

Question-and-answer schema is the raw material AI answer engines quote. Sites with it get lifted into answers; sites without it get summarized vaguely.

AI engines match user questions to headings. Pages whose headings mirror real questions get cited; label-style headings get skipped.

The first substantial paragraph is what AI assistants read to work out what you do. A vague opener produces a vague AI description of your business.

AI systems ingest pages in chunks. Wall-of-text paragraphs over ~90 words often get split mid-thought and lose their meaning.

"Click here" tells Google and AI nothing about the destination. Descriptive links pass meaning — and ranking signal — to your key pages.

This schema is your machine-readable business card: name, logo, contacts. Without it, Google and AI systems piece your identity together from fragments.

sameAs links connect your site to your LinkedIn, socials and directories, proving to machines that all those profiles are one real business.

Both Google and AI engines prefer content they can date. An undated site reads as potentially abandoned.

Google’s quality guidelines reward content with identifiable humans behind it. Anonymous sites are easier to discount.

Local rankings and AI recommendations lean on consistent business name, address, and phone. If they are not on your site, nothing else can corroborate them.

The title tag is the headline of your Google listing. Missing or truncated titles cost clicks even when you do rank.

The meta description is your ad copy in search results. Without one, Google improvises — usually badly.

The H1 tells crawlers what a page is about. Zero H1s leaves them guessing; several H1s blur the topic.

Headings are the outline machines read first. Skipped levels (H1 straight to H4) scramble that outline.

Open Graph controls how your link looks when shared on WhatsApp, LinkedIn and Facebook. Without it, shares show a bare grey link nobody taps.

The card tag upgrades your links on X from plain text to a large image preview.

Pages under ~500 words rarely rank for anything competitive, and give AI engines too little to summarize you accurately.

Internal links are how crawlers find your deeper pages and how ranking strength flows through your site.

A blog is the machinery that earns rankings for the questions customers actually type. Sites without one compete on the homepage alone.

Near-empty pages drag down how ranking systems judge the whole site.

LCP is how long the main content takes to appear on a phone. Google uses it in ranking, and every extra second loses visitors.

INP measures how quickly the page reacts when someone taps. Laggy response reads as "broken" to users.

CLS measures content jumping around while loading — the effect that makes people tap the wrong button and leave.

The HTML document is the first byte of everything. A bloated document delays every other resource behind it.

Below-the-fold images loaded eagerly compete with the content people actually see first.

WebP/AVIF images are 25-50% smaller than JPEG/PNG at the same quality — free speed.

Every blocking stylesheet or script in the head delays the moment anything appears on screen.

Without a viewport tag, phones render the desktop layout shrunk to unreadable size. Google flags the page as not mobile-friendly.

srcset lets phones download phone-sized images instead of desktop ones — often the single biggest mobile data saving.

Alt text is how search engines, AI systems and screen readers see your images. Missing alt text is invisible inventory.

Disabling zoom traps users who need larger text and is penalized in accessibility audits.

Semantic landmarks (<nav>, <footer>, <main>) tell machines which part of the page is content versus chrome — cleaner extraction for search and AI.

HSTS makes browsers refuse ever to load your site insecurely — closing a classic interception hole on public Wi-Fi.

A CSP limits what scripts can run on your pages — the main defence against injected malicious code.

This one-line header stops browsers from "sniffing" files into executable types — a cheap, standard hardening step.

Without frame protection, your site can be invisibly embedded in a hostile page and your visitors tricked into clicking things.

Loading any http:// resource on an https:// page breaks the padlock and gets scripts blocked outright by modern browsers.

A business customers cannot phone or email is a business they do not trust — and ranking systems read the same signal.

Links to active social profiles corroborate that the business is real and current.

The favicon shows next to your name in search results and browser tabs. A missing one reads as unfinished.

A privacy policy is both a legal requirement in most markets and a quality signal ranking systems check for.